Open standards (wandering, draft)
Apr. 22nd, 2017 12:23 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
jptb(The examples used in this article are probably specific to the US.)
I'm a huge fan of open standards. Without them, much of what we take for granted wouldn't be possible. Replaceable parts, APIs, standardized sizes and weights all make our lives simpler. Open standards turn bespoke items into commodities. This generally benefits producers (cheaper, more easily obtained components) and consumers (cheaper, more easily repaired goods).
Open standards can be prevented by patents and copyrights. However, there are limits to what a patent covers. Courts in the US [1] and UK [2] have ruled that a patented device created from unpatented components does not allow the patent holder to restrict others from making/making use of the unpatented components.
In short, if I receive a patent for a new type of electric motor, I can't prevent someone from making replacement parts for it unless I create novel, patentable parts for the entire device. This is generally more of a pain in the rear than it's worth.
For physical goods, generally speaking, we have a right to make and use replacement physical components even if the maker of the good doesn't want us to. Open standards are in the best interest of manufacturers and, because of court rulings, we're allowed to benefit from them as well.
Unfortunately, open standards are actively worked against in software and, as Marc Andreessen noted, software is eating the world.
John Deere, joined by many other firms, claims the software that runs their products is copyrighted and that consumers who purchase a physical good hold only a license to the software that makes it run. [3] This prevents and/or complicates Deere product owners from working on their purchased goods.
Tesla restricts what you can do with your Tesla. They do this based on the license for the software that controls the car, not based on a legal right to tell you what you can do with the physical vehicle. [5] Generally speaking, the maker of a physical good cannot restrict you from using it as you like. For example, Ford can't sell you a car and say you may not make left turns with it.
Oracle is (still) suing Google, claiming Oracle holds copyright on the Java API and that Google has infringed this copyright. [5]
For software, common sense and physical good case law is increasing being thrown out as market leaders work to permanently entrench themselves at the top by restricting the use of "their" "intellectual property". The idea that "makers" have a right to control the use of "their" product post-sale is bad for everyone but the "maker".
Which is why we have to mandate that APIs are not subject to copyright.
Every API is merely a set of method signatures. Is f(list<string, double>) -> double copyrightable? Should it be? Is a set of these signatures copyrightable? Should it be? If so, why? Millions of different functions fit this definition. Should one company control them all?
If APIs are not copyrightable, we can create new implementations that allow us to do what we want with things we own. But we still need the right to install the software.
[1] https://en.wikipedia.org/wiki/Aro_Manufacturing_Co._v._Convertible_Top_Replacement_Co.
[2] https://en.wikipedia.org/wiki/British_Leyland_Motor_Corp_v_Armstrong_Patents_Co
[3] http://www.npr.org/sections/alltechconsidered/2015/08/17/432601480/diy-tractor-repair-runs-afoul-of-copyright-law
[4] http://www.latimes.com/opinion/op-ed/la-oe-perzanowski-schultz-tesla-software-ownership-20161104-story.html
[5] https://www.theregister.co.uk/2017/02/11/oracle_refuses_to_let_java_suit_die/
I'm a huge fan of open standards. Without them, much of what we take for granted wouldn't be possible. Replaceable parts, APIs, standardized sizes and weights all make our lives simpler. Open standards turn bespoke items into commodities. This generally benefits producers (cheaper, more easily obtained components) and consumers (cheaper, more easily repaired goods).
Open standards can be prevented by patents and copyrights. However, there are limits to what a patent covers. Courts in the US [1] and UK [2] have ruled that a patented device created from unpatented components does not allow the patent holder to restrict others from making/making use of the unpatented components.
In short, if I receive a patent for a new type of electric motor, I can't prevent someone from making replacement parts for it unless I create novel, patentable parts for the entire device. This is generally more of a pain in the rear than it's worth.
For physical goods, generally speaking, we have a right to make and use replacement physical components even if the maker of the good doesn't want us to. Open standards are in the best interest of manufacturers and, because of court rulings, we're allowed to benefit from them as well.
Unfortunately, open standards are actively worked against in software and, as Marc Andreessen noted, software is eating the world.
John Deere, joined by many other firms, claims the software that runs their products is copyrighted and that consumers who purchase a physical good hold only a license to the software that makes it run. [3] This prevents and/or complicates Deere product owners from working on their purchased goods.
Tesla restricts what you can do with your Tesla. They do this based on the license for the software that controls the car, not based on a legal right to tell you what you can do with the physical vehicle. [5] Generally speaking, the maker of a physical good cannot restrict you from using it as you like. For example, Ford can't sell you a car and say you may not make left turns with it.
Oracle is (still) suing Google, claiming Oracle holds copyright on the Java API and that Google has infringed this copyright. [5]
For software, common sense and physical good case law is increasing being thrown out as market leaders work to permanently entrench themselves at the top by restricting the use of "their" "intellectual property". The idea that "makers" have a right to control the use of "their" product post-sale is bad for everyone but the "maker".
Which is why we have to mandate that APIs are not subject to copyright.
Every API is merely a set of method signatures. Is f(list<string, double>) -> double copyrightable? Should it be? Is a set of these signatures copyrightable? Should it be? If so, why? Millions of different functions fit this definition. Should one company control them all?
If APIs are not copyrightable, we can create new implementations that allow us to do what we want with things we own. But we still need the right to install the software.
[1] https://en.wikipedia.org/wiki/Aro_Manufacturing_Co._v._Convertible_Top_Replacement_Co.
[2] https://en.wikipedia.org/wiki/British_Leyland_Motor_Corp_v_Armstrong_Patents_Co
[3] http://www.npr.org/sections/alltechconsidered/2015/08/17/432601480/diy-tractor-repair-runs-afoul-of-copyright-law
[4] http://www.latimes.com/opinion/op-ed/la-oe-perzanowski-schultz-tesla-software-ownership-20161104-story.html
[5] https://www.theregister.co.uk/2017/02/11/oracle_refuses_to_let_java_suit_die/
